2 nd Int.

PhD Symposium in Civil Engineering 1998 Budapest

DURACRETE – PROBABILISTIC PERFORMANCE BASED

DURABILITY DESIGN OF CONCRETE STRUCTURES

Anders Lindvall1
Chalmers University of Technology, Department of Building Materials
SE-412 96 Göteborg
Sweden

SUMMARY

The presented project is a first attempt to make probabilistic performance based

durability designs of concrete structures. The performance based durability design is
based upon realistic models that describes the future behaviour of the concrete and the
environment. To make calculations it is necessary to quantify the parameters included in
the chosen models. In this paper two models that describes the behaviour of concrete are
presented and it is shown how the environmental parameters in the models are
statistically quantified.

Keywords: concrete, durability, environmental influence, carbonation, chloride ingress,

statistical quantification

1. INTRODUCTION

The present design with respect to the durability of concrete structures is in large extent
empirical. It is in large extent based on deem-to-satisfy rules, where for example
minimum concrete cover and maximum water/cement ratio is prescribed. With this
approach the structure will have an acceptable long, but not specified lifetime.

If a performance based durability design methodology is used, it is possible to make

more accurate estimations of the lifetime of a concrete structure. Performance based
means in this respect, that the ability to fulfil relevant functions is quantified. The
design methodology is based on realistic and sufficiently realistic environmental and
material models capable to predict the future behaviour of a concrete structure. In the
performance based durability design methodology the lifetime is determined by
probabilistic calculations.

To make probabilistic lifetime calculations it is necessary to find models for the

deterioration of concrete and the structural behaviour of deteriorated concrete. In the
deterioration models it is important to find the influencing factors, for example concrete
properties, curing and the environment, and to quantify them. This is sometimes a hard
task, since the quantification of the parameters are based on observations, made in
laboratory and field-exposure. Often is little known about the background of the
observations, especially from field-observations. This leads to a large scatter when the
parameters in the deterioration-models are quantified.
1
PhD student

1
2. MODELS FOR DEGRADATION

The degradation of concrete is commonly modelled as occurring in two stages defined

as “initiation” and “propagation”. The initiation period is the period during which
changes in the environment within the concrete continue to take place, normally as a
result of interaction with the exposure environment, until, eventually, a limit is reached
when damage occurs. The propagation period begins at the moment a certain defined
event occurs, e.g. corrosion starts, cracks appears etc, until a specified limit state is
reached.

To be able to make predictions of the degradation of reinforced concrete structures there

is a need of models that describe the degradation-processes. When studying the
degradation of concrete three different types of degradation can be identified:
reinforcement corrosion, alkali-aggregate reactions and frost attack. In reality it is only
for reinforcement corrosion that models that can be used for predictions exist.

In the following the two models I have chosen to work with are presented. The
following degradation process are treated:
• Carbonation-Induced Corrosion.
• Chloride-Induced Corrosion.

2.1 Carbonation-Induced Corrosion

One model to described the carbonation-process in concrete was presented by the CEB
Task Group V, 1+2 (1996). The model is mathematically represented as follows:

2 ⋅ k1 ⋅ k 2 ⋅ Deff ⋅ Cs t 
n

xc = ⋅ t ⋅ 0  (1)
a t

where: Deff is the effective diffusion coefficient at a defined execution and

environmental conditions, a is the binding capacity for CO2 , t is the time in service,
t0 is a reference period (e.g. 1 year), k1 is a factor which considers considers the
influence of execution of Deff (e.g. influence of curing), k 2 is a factor which considers
the influence of the environment on Deff (e.g. influence from moisture and temperature
conditions at the surface of the concrete structure) and n is a factor which considers the
influence from the environment on the time-evolution of Deff (e.g. shelter against rain
and moisture conditions at the surface of the structure).

2.2 Chloride Induced Corrosion

One model to describe the chloride ingress-process in concrete was presented by

Mejlbro (1996). The model, rewritten in a convenient from for design purposes, is
mathematically represented as follows:

2
  
 
 x 
Cx = CSN ⋅ 1 − erf n  (2)
  tm 
2 ⋅ Dce( m) ⋅ f c ⋅ f e ⋅ f t ⋅   ⋅t 
  t 

where: Dce ( m) is the effective diffusion coefficient at a defined execution and

environmental conditions measured at time m, n is a factor which considers the
influence from the material and the environment on the time-evolution of Dce ( m) , CSN is
the surface chloride level, t is the exposure period, tm is the reference period (normally
28 days), f c is a factor that considers the influence from curing on Dce ( m) , f e is a factor
that considers the differences between different exposure environments, f t is a factor
that considers the test-method and x is the depth.

3. ENVIRONMENTAL PARAMETERS

The general approach of modelling the environmental actions on concrete structures is

to separate actions from the response of the structure. This means that the “true
environmental actions” in large extent are independent of the choice of concrete. In the
following some examples of environmental actions are given.

3.1 Surface temperature

The best way of describing the surface temperature conditions is to use an equivalent
surface temperature, see (3). In the expression for the equivalent surface temperature the
influences from the temperature of the surrounding air, the solar radiation, convective
heat transfer with the ambient air, long wave radiation exchange with the sky and the
surroundings, evaporation and heat transferred from the inside of the structure are
included.

Ti
I sun ⋅ a + + αr ⋅ Tsky + αcv ⋅ Tair − he ⋅ g
Tsurf = Tair + M (3)
1
+ αr + αcv
M

where: Tair is the air temperature, I sun is the solar radiation, a is the solar radiation
absorption factor, Ti is the temperature inside the structure, M is the heat resistance for
the structure, αr and αcv are the heat transfer coefficients for radiative and convective
heat transfer respectively, Tsky is the sky temperature, he is the evaporation energy and
g is the evaporation rate.

3
3.2 Surface Humidity

Due to solar radiation the surface humidity differs from the humidity in the surrounding
air. Under the assumption of constant vapour content in the air, the surface humidity is
predicted with (4).

RHair ⋅ vs (Tair )
RH surf = (4)
v s (Ts ,eq )

where: RHair is the relative humidity in the air, vs (Tair ) being the vapour content of air
at saturation at a specific temperature, Tair is the air temperature and Teq is the
equivalent surface temperature.

3.3 Time of Wetness

The time of wetness for a surface is a summation of the time of wetness due to rain and
time of wetness due to surface condensation. This is given in (5).

TOW = t wet ,rain + t wet ,cond (5)

where: t wet ,rain is the time of wetness due to rain and t wet ,cond is the time of wetness due
surface condensation.
The time of wetness due to rain and due to surface condensation are however correlated.
• Precipitation. A division is made between vertical and driving rain.
• Surface condensation. If the equivalent surface humidity exceeds 100% surface
condensation will occur, i.e. RH surf ≥ 100% .

3.4 Chloride conditions

The true environmental actions in a chloride environment should be described at the

concrete surface and include temperature, humidity and chloride concentration. A way
should be find to go from these actions to the response by the concrete, e.g. described
by a surface chloride content Cs , if empirical models are used. The equation for the Cs -
value is described in (6).

Ctot′ (Cenv , w / c, type of binder , T ) ⋅ C

Cs = (6)
ρconc

where: Ctot′ (% of binder-weight) is given by the chloride binding isotherm at Cenv and a
temperature of +20 °C . The correction factor for T follows from the Arrhenius equation.
C is the cement-content in the concrete and γ conc is the density of the concrete. The
“equivalent chloride concentration”, Cenv , should be different for a marine or a bridge
environment.

4
4. QUANTIFICATION OF ENVIRONMENTAL PARAMETERS

To be able to make probabilistic lifetime calculations it is necessary to quantify the

parameters in the deterioration models. In the following some examples of how the
environmental parameters in a carbonation model and a chloride ingress model can be
performed.

4.1 Carbonation Induced Corrosion

In the model for carbonation presented above three different environmental factors can
be identified. These are:
♦ k2. Environmental factor. This factor considers the differences between the
conditions in a reference climate and the climate conditions on-site.
♦ n. Age factor. This factor considers the age-dependency of the diffusion
coefficient.
♦ CS. Surface concentration of carbon dioxide. This concentration is normally
assumed to be equal to the concentration of carbon dioxide in the surrounding air
(equal to 358 ppm).

The quantification of the factors is made with data from a 16-year exposure test on 27
concrete mixes and for three different exposure conditions, indoors (LAB), outdoors
sheltered (OS) and outdoors unsheltered (OUS). The results from the exposure test are
presented in Wierig (1984).

The quantification is made in such way that a reference climate is defined. Normally the
indoor climate is used as a reference climate. In this reference climate the environmental
parameter, k2, is equal to 1 and the age factor, n, is equal to 0. The quantification is
made in the following way:
1. The n -parameter is quantified by making a linear regression analysis in a log-log
diagramme, with the carbonation-depth on one axis and the time on the other axis.
2. The k 2 -parameter is quantified by comparing the carbonation-depths in different
climates for the same concrete-mix and the same execution.

The result of the quantification can look like in figure 1a, where the k 2 -factor is plotted
against the surface humidity, and figure 1b, where the n -parameter is plotted against
rain time. Due to lack of field-observation at different climates, data from calculations
made with a micro-level carbanation model, used in CEB TG V/1+2. The model is
presented in CEB Bulletin 238.

5
Figure 1a: The result of the quantification of the environmental parameter, k2, for OPC-
and GGBS-concrete exposed in an OUS-environment.

Figure 1b: The result of the quantification of the age factor, n, for OPC- and GGBS-
concrete exposed in an OUS-environment.

In figure 1a and 1b the result of the quantification of the field observation are given as
mean-values. To each of these mean-values the following statistical distribution
function are attached:
♦ Environmental factor: LogNorm(0.48;0.26) (OPC-concrete) and LogNorm (0.50;
0.12) (GGBS-concrete).

6
♦ Age-factor: Beta(0.283; 0.184; 0; 0.5) (OPC-concrete) and Beta(0.319; 0.124; 0;
0.5) (GGBS-concrete).

4.2 Chloride Induced Corrosion

In the model for chloride ingress presented above three different environmental
parameters can be presented:
♦ fe. Environmental factor. This factor considers the differences between the
conditions in a reference climate and the climate-conditions on-site.
♦ n. Age-factor. This factor considers the age-dependency of the apparent diffusion
coefficient.
♦ CSN. Surface chloride concentration. This factor describes the driving potential of
chlorides into the concrete structure.

The quantification of the parameters is made with data from Nilsson et al (1997), where
15 different concrete mixes are studied in a 2-year exposure programme in a marine
environment on the Swedish west coast. Three different zones in the marine
environment have been studied: the submerged zone (S), the splash zone (Sp) and the
atmospheric zone (A).

The quantification is made in the following steps:

1. The age factor, n, is quantified by making a linear regression analysis in a log-log
diagram, with the apparent diffusion coefficient as a function of time.
2. Using the result from the quantification of the age factor an apparent diffusion
coefficient for 28 days is calculated.
3. The environmental factor, fe, is quantified by comparing the calculated apparent
diffusion coefficient for 28 days, in the different environments, with a diffusion
coefficient achieved in laboratory.
4. The surface chloride concentration, CSN, is quantified by writing the surface
chloride concentration as a function of w/b.

The result of the quantification can look like in figure 2a, where the statistical
distribution functions for n for OPC-concrete in the submerged and splash zones are
given, and in figure 2b, where the statistical distribution functions for the fe for OPC-
concrete in S, Sp and A are given, and in figure 2c, where CSN is given for OPC-
concrete as a function of w/b in S, Sp and A.

7
 Density Plots (2 Graphs) - [Age-fact]
Relative Frequency
15.4730
Submerged
Beta
13.9366 Splash
Beta

12.4003

10.8639

9.3276

7.7913

6.2549

4.7186

3.1822

1.6459

0.1095
0.0 0.1 0.1 0.2 0.2 0.3 0.3 0.3 0.4 0.4 0.5 0.5 0.6 0.7 0.7 0.8 0.8 0.9 0.9 1.0 1.0
Value of X

Figure 2a: The result of the quantification of the age-factor, n, for OPC-concrete
exposed in the submerged, splash and atmospheric zones.
Density Plots (1 Graphs) - [Environm]
Relative Frequency
0.7796
HETEK
Gamma
0.7022

0.6248

0.5474

0.4700

0.3926

0.3151

0.2377

0.1603

0.0829

0.0055
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0
Value of X

Figure 2b: The result of the quantification of the environmental factor, fe, for OPC-
concrete exposed in the submerged zone.

8
Figure 2c: The result of the quantification of the surface chloride concentration, CSN,
for OPC-concrete exposed in the submerged, splash and atmospheric zones.

5. PROJECTS

In the last couple of years there have been some projects, where this new probabilistic
approach to durability problems has been applied. On the project, where the author of
this paper has been involved, is DuraCrete2.

5.1 Duracrete

DuraCrete is a European project financed by the European Union. The project is a joint
project between 12 different partners, companies, institutes and universities, in the EU-
member states. The project is divided into eight different sub-tasks. In each sub-task
between two and eight different partners are involved.

Within the DURACRETE-project calculations will be made where the lifetime for a
number of different concrete structures will be estimated. The input data for the designs
are achieved from the quantification made in the project. The quantification is based on
observation made in the project and found in literature.

The end of the lifetime is decided by the chosen limit-state. A limit-state is for example
when the reinforcement starts to corrode, when cracks occur or when the structure
collapse due to corroding reinforcement. This means if for example the last limit-state is
selected the limit-state function has to include everything that happens with the
structure until the collapse. Consequently the limit-state function has to include the

2
Brite-Euram Project BE95-1347: DuraCrete - Probabilistic Performance Based Design of Concrete
Structures. Partners: COWI (DK), CUR (NL) (coord.), CTH (S), Geocisa (E), HBG (NL), IBAC (D),
IETCC (E), Intron (NL), Schwenk (D), RWS (NL), Taywood (UK) and TNO (NL).

9
initiation and propagation of corrosion on the reinforcement and the structural
consequences of the corrosion.

6. REFERENCES
Mejlbro, L.: The complete solution of Fick’s second law of diffusion with time-
dependent diffusion coefficient and surface concentration, Durability of Concrete in
Saline Environment, Cementa AB, Danderyd Sweden, 1996. pp 127-158.
Nevander, L.-E. & Elmarsson, B.: Fukthandbok - praktik och teori (Moisture
handbook - practice and theory), Svensk Byggtjänst, Stockholm, 1992. (in Swedish).
Nilsson, L.-O. & Sandberg, P. & Poulsen, E. & Tang, L. & Andersen, A. &
Fredriksen, J.M.: HETEK, A system for estimation of chloride ingress into concrete,
Theoretical background, Danish Road Directorate, Report 83, Københaven, 1997.
Nilsson, L-O., & Rodhe, M.: The CTH carbonation model - a micro-level model for
carbonation in a natural climate, New approach to durability design, CEB Bulletin No.
238, Lausanne, 1997. Appendix 2.
Wierig, H-J.: Long-time studies on the carbonation of concrete under normal outdoor
exposure, Proceedings of the RILEM seminar on Durability of concrete structures
under normal outdoor exposure, 26th-29th March 1994, Hannover, 1994. pp 239-249.

10